Yes, but;
These programs by John Broskie "Glassware" are specifically made for tube circuits and you can choose 51 different topologies to begin with, that will give the results you need. Now, considering IC's, Different biasing topology, and the like, Yes it is limited in that regard. I do like LTSpice design parameters window, because it enables you to see the losses and recommends ALL COMPONENTS that would fit criteria that would enable one to choose as parts specifically of a certain value may not be obtainable because of not being manufactured. So, it would be great If someone had all of these programs, When designing Tube Amplifiers Like the one on this thread site the "EZ10" model. Pianolydia
These programs by John Broskie "Glassware" are specifically made for tube circuits and you can choose 51 different topologies to begin with, that will give the results you need. Now, considering IC's, Different biasing topology, and the like, Yes it is limited in that regard. I do like LTSpice design parameters window, because it enables you to see the losses and recommends ALL COMPONENTS that would fit criteria that would enable one to choose as parts specifically of a certain value may not be obtainable because of not being manufactured. So, it would be great If someone had all of these programs, When designing Tube Amplifiers Like the one on this thread site the "EZ10" model. Pianolydia
The schematic has a 10 watt GXPP10-8-8K transformer, that's why I was wondering (in addition to Eli Duttmann's comments).
I have used LTSpice earlier, and I kind of like it. However, are there tube libraries for it?
Last edited:
Joku -
I did use a 15 watt Edcor output transformer in my buildnot thje 10watt as indicated. They were the 8K version, I had not tried the ones Eli had mentioned (7.6k at 25w). In the original build thread I had ask at one point about that but never got a direct response as to best UL transformer value. I looked at what others were doing and took a best guess.
Sandy
I did use a 15 watt Edcor output transformer in my buildnot thje 10watt as indicated. They were the 8K version, I had not tried the ones Eli had mentioned (7.6k at 25w). In the original build thread I had ask at one point about that but never got a direct response as to best UL transformer value. I looked at what others were doing and took a best guess.
Sandy
Few remarks;
Do I sense that there is resistance to NFB? Why? The facts regarding this have been adequately treated in many other places, not to start a flame war here. Only to venture: I hope one can get rid of some snake oil anecdotal tales regarding the use of nfb .....
About UL: Also often discussed elsewhere. UL combines the advantages of pentodes with the advantages of triodes in an unexpected 'free lunch'. Again some anecdotal 'wisdom' is to be found, but hopefully folks will study some literature if in doubt.
The OP indicated 15W desired with 6V6s. One will be very hard put to get 15W after the output transformer from anywhere near safely used 6V6s. Depending on the prices in various countries, it is very easy to get 15W from a pair of 5881s, 6L6s or whatever equivalent.
Then lastly, there is such a thing as 'music power' (an U.S.A. term). I find that a useful description for a condition equating fixed bias performance with cathode bias. To eliminate a fixed bias supply I would suggest that. These days, with relatively small physical component size, it is practical to use a >1000µF cathode bias bypass capacitor for the power valves, even with a common cathode resistor (I use 2200µF.) That way one retains the long-term self-compensating nature of cathode bias with the performance of fixed bias for 90% of the time. (A large value of C keeps the charge pretty much constant most of the time, closely resembling fixed bias.)
Do I sense that there is resistance to NFB? Why? The facts regarding this have been adequately treated in many other places, not to start a flame war here. Only to venture: I hope one can get rid of some snake oil anecdotal tales regarding the use of nfb .....
About UL: Also often discussed elsewhere. UL combines the advantages of pentodes with the advantages of triodes in an unexpected 'free lunch'. Again some anecdotal 'wisdom' is to be found, but hopefully folks will study some literature if in doubt.
The OP indicated 15W desired with 6V6s. One will be very hard put to get 15W after the output transformer from anywhere near safely used 6V6s. Depending on the prices in various countries, it is very easy to get 15W from a pair of 5881s, 6L6s or whatever equivalent.
Then lastly, there is such a thing as 'music power' (an U.S.A. term). I find that a useful description for a condition equating fixed bias performance with cathode bias. To eliminate a fixed bias supply I would suggest that. These days, with relatively small physical component size, it is practical to use a >1000µF cathode bias bypass capacitor for the power valves, even with a common cathode resistor (I use 2200µF.) That way one retains the long-term self-compensating nature of cathode bias with the performance of fixed bias for 90% of the time. (A large value of C keeps the charge pretty much constant most of the time, closely resembling fixed bias.)
Nah, there's nothing wrong with NFB. We were just talking about having to have magnetic headroom in transformer to cope with lower frequencies. Same thing with UL mode.
It doesn't need to be exactly 15W, 12-14 watts is OK.
I think I'll go with "El Cheapo", but it probably doesn't matter that much as the trafos should work with other 6V6 designs.
Just a bit more dumb questions, what's the cathode voltage and bias current on the 6V6 output tubes? It's a while since I have done stuff like this..
Also, what's the current in the current sink? I think I'll use some other CCS than the one in schematic as the FETs seem to be hard to find. I was suggested to use a LED biased cascoded BJT CCS, any ideas where to find info about those?
Just a bit more dumb questions, what's the cathode voltage and bias current on the 6V6 output tubes? It's a while since I have done stuff like this..
Also, what's the current in the current sink? I think I'll use some other CCS than the one in schematic as the FETs seem to be hard to find. I was suggested to use a LED biased cascoded BJT CCS, any ideas where to find info about those?
Joku -
I just got the 5881 amp running, I'll post an update on that thread, but amp made a touch over 20watts out of the box and I'll say sounds amazing.
Sandy
I just got the 5881 amp running, I'll post an update on that thread, but amp made a touch over 20watts out of the box and I'll say sounds amazing.
Sandy
According to GE data: Ua = 285 V, Ug1 = -19 V, Ia = 35 mA.
I have seen higher anode voltage values, up to 320 V, used with good results.
Usually the output tubes should not be biased to max. allowed Pd (12 W).
The above recommendation of GE gives Pd = 10 W.
I have simply adjusted all CCS's I have used to minimum THD, including the CCS of LTP .
Suitable CCS for a LTP is made of LM317 voltage regulator.
Its voltage rating is (only) 40 V, but still sufficient for this purpose.
LM317 can be used as a CCS for LTP like this, for example:
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Most of the time, that could be true. However, "El Cheapo" applies NFB to the non-inverting triode's grid. In this situation, the tail should be infinitely long. Infinitely long is not possible, but VERY long is. Use a cascoded CCS in "El Cheapo" builds. PCBs for cascoded BJTs were (are?) available from EU based members. Nose around.
The 12AT7/ECC81 sounds good with 200 to 220 V. on the plate and an IB of 3 mA. A B+ rail of 355 V. is "perfect". Set the CCS in each channel at 6 mA. Even if UL mode "finals" are employed, keep the current limiting resistors in the g2 circuitry. Excessive plate and g2 dissipation is what damages tubes, not a few volts over a documented limit.
Last edited:
If I remember right, the dynamic resistance of such LM317 CCS is some 1M.
How much longer tail you see necessary ?
How much longer tail you see necessary ?
I can make a PCB, so I would just need the schematic. 🙂
Also, I did some simulations and the bias current seems to be close to 35mA per tube. With 350V on the plate and ~24V on cathode, this results in voltage differential of ~326V over the tube and about 11.4 watts of dissipation. Is it really OK to be so close to thermal limits of the tube?
Last edited:
In this case, I would say yes. Look at the GE data sheet. Please notice that the 12 W. plate dissipation limit is design center. Design center specifications are quite conservative and operation at approx. 10% above the value given is not unreasonable. IMO, tubes should not be operated at an absolute maximum dissipation, but design maximum usually works out OK. Design maximum is roughly 10% greater than design center.
IMO, you will be fine dissipating 11.4 W./plate. Make damned sure to keep g2 dissipation under control.
Yes - 1M'ish until the device capacitance starts to have influence at higher frequencies. If you do the checks you will see that this happens at way too low a frequency for my liking and so I'm with Eli on this, a cascode BJT is far superior and changing from an LM317 to cascode BJT is quite noticable sonically. Keep the wiring short and sweet. Even a few pF stray capacitance can make an audible difference.
Cheers,
Ian
Last edited:
There is a kit 4 from old Audio Note lineup .All the bad, illogical things; floating paraphase splitter 12ax to 6SN7 driver to a pair of 6V6 in triode mode to AN TRans-200 (OTO transformers ) and feedback as well. One of the most musical PP amps I've heard. Bayshlag resistors , ERo green MKT 0.47 coupling caps, SS Power Supply. Sometimes you got to wonder when superior sounding circuits somehow wear on you quickly and plain vanilla Audio Note keeps making music . I forgot about that simple fact.
Regards, L
Regards, L
At the moment I am unable to check this feature, but if you have any test results or link to such, I would be interested to see.
Can't find it now but one datasheet I saw suggested 15pF output capacitance for the LM317T (TO-220 Version).
Impedance of 15pF is down to 100kOhms by 100kHz.
Cheers,
Ian
Impedance of 15pF is down to 100kOhms by 100kHz.
Cheers,
Ian
- Status
- Not open for further replies.